Design Details & Modifications
Design Details
The Aerocanard's design can be traced directly back to the Rutan Long-Ez, which first flew in 1979.
In the late 1980s Nat Puffer modified the design to create the Cozy, which was very close to a Long-Ez, but with side-by-side seating in the front and a single seat in the rear.
In the early 1990s, this was further modified to create the Cozy MkIV, which was wider in the rear to accommodate two people in the back seat. There were also many other changes including greater wing span, thicker spars, a wider fuselage, etc. Nat was very conservative with the design and published limits, resulting in a very robust aircraft that gives very few problems if handled as specified.
A few years later, the design was forked to create the Aerocanard, which was made available to be built from plans, or as a kit. This could be built to the same dimensions as a Cozy IV, or with a wider canopy and wider fuselage at the firewall, providing significantly more headroom and eliminating the need for bulges on the aft fuselage sides to cover the cylinder heads of the Lycoming engine.

The structure of the aircraft is created from a 'sandwich' of fibreglass and epoxy laid up over foam cores, creating a very lightweight but strong and stiff structure. Using this method, almost any shape can be easily fabricated by first shaping the foam to the desired shape, then laying up fibreglass on top. Compound curves and other complex shapes that are very difficult to achieve in metal are simple in fibreglass.

The most notable features of the design are the canard arrangement (with the small wing in front instead of behind) and the pusher engine, rather than the common engine in front.
The canard layout creates an aircraft that can not stall in the conventional sense. As it is designed so that the canard always stalls before the main wing, the effect is that the nose lowers before the main wing reaches its critical angle, thus a properly built and operated canard aircraft is impervious to the stall-spin accidents that are all too common in conventional designs.
Modifications
Reduced Lower Winglets
Many of these aircraft are flying with no lower winglets at all. However, they experience wing-rock at high angles of attack (near the stall) and possibly a slightly reduced deep-stall margin. The intent is to reduce drag; however, I have yet to see evidence that there is a measurable improvement. As far as I can tell it's a wash, my theory is that the reduced wetted area reduces drag, but slightly increased tip vortex eliminates the gain.
I will have reduced-size lower winglets, but they will still be there. They also protect the winglet attach layups in the event the tips drag, say during a crosswind landing, though this is unlikely with the wide retracts I will have. Perhaps they will protect the wing structure in the event of a wheels-up landing!

Higher Canopy
I'm tall. I want to have comfortable seat cushions under me for comfort on long flights. I don't like bumping my head on the canopy, as I have experienced in many light aircraft (designed by little people, I presume). This is an approved mod. I plan to taper to standard height at the firewall (so the height increase for the back-seaters in relatively small). I will add probably 2", which should be plenty.
Increasing size increases drag, though in turbulent flow (which this area is - expecting laminar flow this far aft would be very optimistic!) the difference will be small.

Wider Canopy
Having to hold your head on an angle to avoid bumping the canopy sucks. My canopy will be 'full-width', that is,  it will extend to the outer edge of the fuselage, rather than the plans inner edge. The canopy will also be more 'square' (not a half-circle) to ensure plenty of headroom.
This change increases frontal area slightly, though as with the higher canopy, the difference in drag should be negligible. It will also eliminate the fuselage 'edge' around the canopy sides, and the extra angles that introduces. So the wetted are may actually be REDUCED, moving towards a reduction in drag, instead of an increase.

Forward-Hinged Canopy
The plans canopy hinges to the side. There are two main issues with this method: 1) the side is curved, but the hinge must be straight, so the hinges stick out into the airflow slightly, and 2) if the canopy becomes unlatched, it may swing open in-flight. Not only would this create a huge distraction, it may release any loose items in the cockpit into the propeller, which could severely damage it.
By having the canopy hinge from the front, if it becomes unlatched in flight, it will merely open very slightly, and it will be a bit windier inside. It can be safely re-latched. And the hinges can be entirely internal, so a small drag reduction, though a small increase in weight is possible.

'Gull-Wing' Rear Cockpit Door
Some front-hinged canopies hinge the entire plans section up. I will forward hinge only the canopy for the front seats and have a separate 'gull-wing' door on one side for the back seaters. I intend at this time to have a longer door than most others have used with this mod, incorporating one long window in it (instead of the stock two side windows, necessary because the plans canopy line goes between them).
As an additional advantage, there will be fixed structure up one side and across between the front canopy and rear door, providing better roll-over protection and additional fuselage stiffness (not that the latter is lacking).
The latching mechanism for the front canopy will be designed to require the rear door to be closed before latching can succeed.

Extended Strakes
A popular mod is to use 'Long-Ez strakes' or 'Cozy Girrrl strakes' in which the inboard section of the strake kinks forward to meet the fuselage near the instrument panel. This provides additional elbow and storage room in the strake roots, with no measurable disadvantages.
My plan at this stage (pending some further discussion and modelling) is to extend the strakes to just forward of the instrument panel, but instead of a big kink, have a straight leading edge all the way to the strake tip, or a smaller kink to increase area further outboard. This will provide two advantages: 1) Wiring and hydraulic lines can be run straight (or nearly so) along the leading edge from ahead of the instrument panel to the strake tips, reducing the wire and line length needed for the equipment outboard; and 2) Increases strake volume, and thus fuel capacity, in-line and slightly ahead of the expected centre of gravity.
The possible problem is the increased area ahead of the centre of gravity, and thus deep-stall susceptibility. Though I am further researching this still, I believe it will be effectively offset in two ways: 1) The sharper sweep angle will reduce the lift-curve of the strake, so the strake lift will not increase linearly with increased area; and 2) As I am building a slightly longer fuselage (see below), the centre of gravity will be further forward anyway, putting more strake area behind the centre of gravity.

'Fat' Strakes
The plans strakes have the same (or very similar) leading and trailing edge shape and thickness as the main wing where it joins to the strake tip. The additional length is simply a straight line, and flat area between the leading and trailing edge shapes.
Instead, I will add a small amount of additional thickness to the strake roots, such that the top of the strakes reach nearly to the top of the upper longerons, and the lower strake skins about 1" - 2" deeper. Thus the upper and lower skins will not have any flat areas, they will be curved from leading edge to trailing edge. This will increase strake volume further, increasing fuel and storage capacity slightly. There should be little or no increase in drag, as the air will be following a smooth, curved path. It worked on the Beech Starship!

'Original Length' Canard
The original Cozy MkIV canard length was reduced to increase deep stall margin by reducing the amount of lift that could be achieved by the canard. It works, but as a consequence front-seat loading capability is reduced, and take-off roll may be increased due to the reduced canard / elevator authority.
I will build to the original length, and mitigate deep stall by other means. As with most of these modifications, thorough flight testing will be required to prove this is safe.

Retractable Main Gear
I will be installing Infinity retractable gear . Retracting this main gear will reduce drag, and provide superior ground handling due to the wider stance and oleo suspension. There is also a safety benefit, as it enables an emergency, wheels-up landing should one be required, such as a water landing or landing on soft terrain. In both of these scenarios, the fixed gear can cause the front end of the aircraft to suddenly 'dig-in' causing a very sudden stop, and/or the fixed gear may be torn out of the fuselage immediately behind the back seat.
Fuel capacity in the strakes is reduced due to the gear wells, but the space for the fixed gear is freed up for a sump tank, replacing some of that lost capacity. The rest of the fuel volume loss will be more than compensated for with the other strake modifications listed on this page.

Better Brakes
Lightweight Berringers would be nice, but at least the 3-puck Matcos will be installed. The plans brakes are barely adequate, and if hot, may not provide the required stopping power. Improved brakes are a recommended mod.

Hydraulic Nose Gear
Since there will be a hydraulic system installed to actuate the main gear, it should save weight to add a few hoses, valves and an actuator to the nose gear rather than add a heavy, dedicated electric retract system.

Hydraulic Landing Brake
I probably won't do this. With the retractable gear extended, drag is increased considerably, making a landing brake unnecessary. Eliminating it will reduce weight and build time.
But, as an option, a landing brake could take advantage of the existing hydraulic system. Would likely enable the use of the landing brake at higher speeds, though it's structural capability at higher speeds will need to be considered.

Fixed Landing Lights (in wingtips)
Per plans, retractable landing lights are mounted under the seats. Fixed lights inside the wingtips will be lighter and able to be used as flashing recognition lights, as well.

Hanging Rudder Pedals
Keeps the floor clear for resting feet, and the mechanism away from possible jams.

Remove Cockpit Access Door
Per plans, a small hatch in the side of the fuselage enables access to the canopy latch. Instead, a flush or streamlined handle will provide this function, without the large, leak-prone door.

Nose Wheel Doors
Provides a smooth, low-drag belly. Reduction in the boundary layer thickness if a belly scoop is used for engine cooling is a possible benefit.

Electric Pitch Trim
I may use this instead of the plans manual system IF further research confirms it has an operational benefit and does not add much weight, which I believe is the case.

Vortex Generators
On both the canard and main wing, reduces stall speed (and thus landing and take-off speed) which is a safety and operational benefit. Should also reduce any effect of dirty leading edges (bug splatter).

Wider Fuselage
I don't like having my shoulders pressed against those of my passenger. The extra width should minimise or eliminate this, as well as provide extra comfort and wider panel area.
The widening is 6" at the front seatback, and reduces to stock Aerocanard FG width at the firewall. Forward of the front seat, the fuselage tapers smoothly down to meet the original profile a few inches ahead of F22, creating a smooth, bullet-like profile. The required changes to the bulkheads to achieve this smooth taper was determined using a CAD model, from which I printed new templates.

The possible problem with a wider fuselage is that the aircraft becomes more prone to 'deep-stall' due to the increased area ahead of the centre of gravity. The taper of the widening should reduce this somewhat, and the addition of 6" to the centre section spar will increase area aft of the CoG. Revised CoG range and / or aerodynamic tweaks can take care of the rest, to be proven in flight test.

Longer Fuselage
I'm above-average height. My children, by the time this machine is flying, will likely be tall as well. When I had the opportunity to sit in a Cozy III,  the length available was really marginal, and that was with NO seat cushions installed. The original designer of the Cozy approved moving the front seatback aft 1", and then reduce or remove the seat cushions as needed for more length! I had two issues with that method: 1) I would have to steal space from the back seat passengers (see aforementioned children) when that space is fairly small already, and 2) I really don't want to sacrifice comfort (in the form of seat cushions) when I am intending to fly this bird long-ranges.
I have added 6" to the cockpit length, 3" for the front seats and 3" for the rear seats. Combined with enlarged leg-holes in the instrument panel, I should have a comfortable amount of length available AND comfortable seats. Plus, the space and utility of the back seats is improved.
The nose is a few inches longer as well, to give a pointier (but not POINTED) less blunted shape. The profile has been drawn in CAD to create a smooth shape and ensure the fuselage bulkheads fit nicely.

The big issue in doing this is that the plans centre of gravity range becomes irrelevant, and a new range will need to be determined. As a rough initial estimate, a Cozy that was made 12" longer (in the main body, not the nose) ended up with it's CoG range moved forward 1/3 of the extension, which makes sense given the lift contribution of the canard is approximately 30%. Of course, the final CoG range will need to be tested and confirmed by flight-test.

Baggage Pods
Probably to be added after initial flight testing is complete. Baggage pods are a popular accessory for these aircraft and can be bought as an easy to complete kit. I'll probably design and build my own, probably as conformal pods / aux tanks (like Dick Rutan used) rather than the now popular hanging type.

Auxiliary Fuel Tanks
For very long-range flights, auxiliary tanks will be necessary. Conformal external tanks, interchangeable with baggage 'pods' are one option, and back-seat auxiliary tanks feeding the sump tank are another. One of these or a combination of the two should enable the maximum range needed, without having to cruise at lower speeds.

Range Extenders
For the crew, not the plane.  Design TBD. A fixed-gear Cozy can run vent lines down the gear legs to outlets near the wheels, but I don't have that exact option. Maybe a vent in a low-pressure area on the aft-end, perhaps a small protruding pipe at the trailing edge where the cowl and wing meet.

Longer Ailerons / Wing Fences
To improve roll rate, and -particularly with the fences- roll-response at low speed. Simulations indicate that lengthening the ailerons provides a more significant improvement in roll performance than deeper ailerons. This has been implemented, apparently with good success, in the past. 
The wing fences improve roll response at low speeds and higher angles of attack when spanwise flow becomes more significant. Control freak? Sure!

Carbon-Fibre Control Surfaces / Cowls
Still researching this one, but the use of a properly selected carbon rather than fibreglass on the flight controls will yield a lighter, stiffer surface that is easier to balance and provide greater flutter resistance to a higher speed.
A carbon-fibre cowl can be built lighter and stiffer than a glass one. You can't buy one off-the-shelf, but since I will need to build a custom cowl anyway (due to the non-plans engine and firewall shape), why not make it as light as possible?
Per-plans, no carbon is used on the aircraft, but there is some low-hanging fruit where the use of carbon in selected areas can provide improvements without great cost (carbon is EXPENSIVE). There are some that are using carbon more extensively in the structure of the aircraft. However, there seems to be little to gain by doing this without re-engineering the whole structure, which is well beyond the scope of the modifications here. If I was going to do that, I'd be building something entirely new.

Liquid-Cooled Engine
There are more and more aircraft flying with auto-conversions with good success. The old advice to stick to the plans and just bolt on a Lycoming is good advice... but I'm clearly not good at sticking to the plans. And I don't have the eternal confidence in Lycomings that some people do. I've not had one fail, but a few have threatened me!
What I am NOT intending to do is to break entirely new ground and do something no-one has done before.
I will use an engine that has been proven effective and reliable in other aircraft before, rather than re-inventing the wheel. That's not to say it will be a bolt-on system, however. What I am looking for is a modern, 6+ cylinder engine that can be turbo-normalized, and is liquid-cooled. Exactly which engine I will use will be determined closer to the time that I need the engine, but an example of what I am considering is a Honda or Suzuki V6, common on Titan Mustangs. Some of these are supercharged as well, and the 3rd-party ECU and gearbox are well proven. Rotaries are very interesting, however, unless there is a new, improved type available, the higher fuel consumption (yes I know, comparable to Lycosaurs, but I am trying to go better than that) and highly experimental nature of these installations reduces their appeal. As I am not a professional engineer, I will probably employ experienced specialists to assist in the (minor) modifications, integration, and testing of the powerplant.
Some of the benefits of using a modern converted engine are better efficiency, lower vibration, easy starting, ability to run auto fuels (with caution), and economic overhaul. Initial costs will not be less than a good used Lycoming; I am not taking this route to save money on the initial installation.
Cooling the engine is another benefit. With proper ducting on both sides of the radiators, cooling drag can be minimized. With the use of eductor-type exhaust, ground-cooling may be greatly improved.
​Oh, and MORE POWER.